palestine polytechnic university deanship of graduate

Palestine Polytechnic University

Deanship of Graduate Studies and Scientific Research

Master of Electrical Engineering

Voltage Profile Improvement Using DSTATCOM Based on

Artificial Intelligent Techniques

Submitted By:

Zaher Abdullah Saafin

Supervisor:

Dr. Fouad R. Zaro

In Partial Fulfillment of the Requirements for the Degree of Master

of Electrical Engineering

Jan 2019

The undersigned hereby certify that they have read and recommend to the Dean-

ship of Graduate Studies and Scientific Research at Palestine Polytechnic University

the acceptance of a thesis entitled:.

Voltage Profile Improvement Using DSTATCOM Based on Artificial In-

telligent Techniques

Submitted by Zaher Abdullah Saafin

In Partial Fulfillment of the Requirements for the Degree of Master of Electrical

Engineering

Graduate Advisory Committee Members:

Dr. Fouad R. Zaro Palestine Polytechnic University Signature:..........................Supervisor

Dr. Nassim Iqteit Palestine Polytechnic University Signature:..........................Internal Examiner

Tamer Khatib An-Najah National University Signature:..........................External Examiner

Approved for the Deanship:

Dean of Graduate Studies and Scientific Research - Palestine Polytechnic University

Signature:................................................ Date:........................

Abstract

Recently the demand of electric power is increased due increasing the residential

and industrial facilities which may contain sensitive nonlinear loads that needed

high power quality (PQ) on distribution system to avoid malfunction operation.

One main PQ issue is voltage profile improvement with acceptable voltage har-

monic distortion. It should be regulated to be within acceptable standard levels.

In order to improve the voltage profile, the distribution static synchronous compen-

sator (DSTATCOM) is used with developed control strategy.

In this research DSTATCOM control is developed based on artificial intelligent (AI)

using artificial neural network(ANN), which depends on optimum values obtained

by using particle swarm optimization (PSO). The results of simulation proved the

superiority and robustness of the proposed control strategy of DSTATCOM for im-

proving the voltage profile on distribution system. The validation of results have

been done by MATLAB/Simulink software package.

ii

iii

jÊÓ

Q¢. éK. Õºj

JÖÏ @

ñªÖÏ @ Ð@Y

j

JAK. ù

KAK. QêºË@ É

®

JË @ ÐA

¢

úÎ« Yêm.

Ì'@ám

'

ú«A

J¢B@ ZA¿

YË@

á¯Aª é<Ë @YJ.« Që@ P

éJ

KAK. QêºË@

éY

JêË @ Q

Jk. AÓ

¬@Qå

A K.

ðP QË @ X@

ñ¯. X

ú

æË @

éJ

JºË@ð

éJ«A

JË@

¯@QÖÏ @

èXAK

Péj. J

K

éJ

KAK. QêºË@

é¯A¢Ë@ úÎ« I. Ê¢Ë@ YK@ Q

K A

JKYg

éJËA«

èXñm.

'.

éJ

KAK. Qê»

é¯A£ h. A

Jm

' ú

æË @ð

éAk

éJ¢

k Q

« ÈAÔg@ úÎ« ø

ñ

Jm

'

à@ áºÜØ

. AêÊÔ« ú

¯ ÉÊ

g

HYm'

B úæk

É®

JË @ ÐA

¢

úÎ«

éËñJ.

®Ó

HAJ

®

¯ @ñ

JK. ù

KAK. QêºË@ Yêm.

Ì'@ám

'

é¯A¢Ë@

èXñk.

HC

ªÓ Yg@

ú

¯

ñªÖÏ @

JJ.¢

Õç

' ½Ë

YË ,

ékñÒÖÏ @ QKAªÖÏ @ áÖÞ

àñºJË Yêm.Ì'@ Õæ

¢

JK I. m.

' . ù

KAK. QêºË@

ZA¿YË @

Q£ úÎ« YÒ

Jª

K

ñªÖÏ @ ú

¯ Õºj

JÊË

é®KQ£

JJ.¢

ð h@Q

¯@ Õç

'

é@PYË@ è

Yë

AîDÊ« ÈñmÌ'@ Õç'

éJËA

JÓ Õæ

¯ úÎ«

èYÒ

JªÖÏ @

éJ. ªË@

HA¾J.

Ë@

éJÓ

PP@ñ

k Ð@Y

jJAK. ú

«A

J¢B@

.éJ

KAK. Qê» ©K

PñK

éºJ.

úÎ« Aê

®JJ.¢

ð PñJ¢Ë@ H. @Qå @

é»Qk áÓ úkñ

JÓ

éJÓ

PP@ñ

k

JJ.¢ áÓ

Yêm.Ì'@

ájJË

ñªÖÏ @ ú

¯ Õºj

JË @ ú

¯

ékQ

®ÖÏ @

éJj.

K @Q

JB@

èñ

¯

èA¿ AjÖÏ @ l .

'A

JK

HQê

£

@

. H. CJÖÏ @

HAJm.

×QK. ÈC

g áÓ l .

'A

JJË @ áÓ

®j

JË @ Õç

' . ©K

PñJË @

éºJ.

úÎ«

DECLARATION

I declare that the Master Thesis entitled ”Voltage Profile Improvement Using

DSTATCOM Based on Artificial Intelligent Techniques” is my own original work,

and hereby certify that unless stated, all work contained within this thesis is my

own independent research and has not been submitted for the award of any other

degree at any institution, except where due acknowledgement is made in the text.


Signature:.................................. Date: .........................

iv

STATEMENT OF PERMISSION

TO USE

In presenting this thesis in partial fulfilment of the requirements for the master

degree of Electrical Engineering at Palestine Polytechnic University, I agree that

the library shall make it available to borrowers under rules of the library. Brief

quotations from this thesis are allowable without special permission, provided

that accurate acknowledgement of the source is made. Permission for extensive

quotation from, reproduction, or publication of this thesis may be granted by my

main supervisor, or in his absence, by the Dean of Graduate Studies and Scientific

Research when, in the opinion of either, the proposed use of the material is for

scholarly purposes. Any copying or use of the material in this thesis for financial

gain shall not be allowed without my written permission.


Signature:.................................. Date: .........................

v

DEDICATION

This work is dedicated to my Parents, my wife, my lovely sons, my lovely daughters,

my brothers and my friends During the master journey they continued to support

and encourage me until I accomplished this work.

vi

ACKNOWLEDGEMENT

I would like to express my deep and sincere gratitude to my supervisor Dr. Fouad

R. Zaro from Palestine Polytechnic University (PPU) for the useful comments,

remarks and support.

Finally, I would like to thank Palestine Polytechnic University for giving me

the opportunity to complete my master’s degree during my work time.

vii

KEYWORDS

Artificial Neural Networks (ANNs), DSTATCOM, Particle Swarm Optimiza-

tion(PSO), PI controller, Power Quality (PQ), Smart Grid (SG), Total Harmonics

Distortion(THD), Voltage Profile.

viii

Contents

1 Introduction 1

1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Thesis Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Thesis Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Literature Review 5

2.1 PQ Problems in Power System . . . . . . . . . . . . . . . . . . . . . 5

2.2 Voltage Profile Improvement Using DSTATCOM . . . . . . . . . . . 6

2.3 Control Strategies of DSTATCOM . . . . . . . . . . . . . . . . . . . 7

2.4 Merging RES with DSTATCOM in SG . . . . . . . . . . . . . . . . . 8

2.5 Tuning DSTATCOM Controller Using AI Techniques and Optimiza-

tion Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.6 Research Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Power Quality Factors 13

3.1 PQ Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

ix

CONTENTS x

3.2 Importance of PQ Factors in Power System . . . . . . . . . . . . . . 13

3.3 Voltage Events of PQ . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3.3.1 Classifications Voltage Events . . . . . . . . . . . . . . . . . . 14

3.3.2 Measurement of Duration and Amplitude of Voltage Event . . 16

3.4 PQ Problems Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 DSTATCOM Configuration and Design 20

4.1 Principle and Operation of DSTATCOM . . . . . . . . . . . . . . . . 20

4.1.1 Equivalent Model of DSTATCOM . . . . . . . . . . . . . . . . 20

4.1.2 Principle of Operation . . . . . . . . . . . . . . . . . . . . . . 23

4.1.3 STATCOM V-I Characteristic . . . . . . . . . . . . . . . . . . 24

4.2 DSTATCOM Control Strategy . . . . . . . . . . . . . . . . . . . . . . 27

4.3 System Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . 28

4.3.1 Load Model Mathematical Equations . . . . . . . . . . . . . . 28

4.3.2 VSI AC-Side Model . . . . . . . . . . . . . . . . . . . . . . . . 29

4.3.3 VSI DC-Side Model Equations . . . . . . . . . . . . . . . . . . 30

4.4 DSTATCOM Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.4.1 Design DC Bus Voltage . . . . . . . . . . . . . . . . . . . . . 31

4.4.2 Design DC Bus Capacitor . . . . . . . . . . . . . . . . . . . . 31

4.4.3 Design AC Interfacing Inductor(Lf ) . . . . . . . . . . . . . . . 32

4.4.4 Design Passive High Pass Ripple Filters . . . . . . . . . . . . 32

CONTENTS xi

5 Methodology 33

5.1 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

5.2 Calculation of DSTATCOM Configurations . . . . . . . . . . . . . . . 34

5.3 MATLAB Model Simulation . . . . . . . . . . . . . . . . . . . . . . . 35

5.3.1 MATLAB Simulink Simulation . . . . . . . . . . . . . . . . . 35

5.3.2 MATLAB Simlink For DSTATCOM . . . . . . . . . . . . . . 35

5.3.2.1 DSTATCOM Controller Simulation . . . . . . . . . . 36

5.4 Artificial Intelligence and Optimization Techniques . . . . . . . . . . 38

5.4.1 Particle Swarm Optimization . . . . . . . . . . . . . . . . . . 39

5.4.1.1 PSO Topology . . . . . . . . . . . . . . . . . . . . . 39

5.4.1.2 Tuning PI Controller Gains of DSTATCOM Using

PSO . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

5.4.2 Artificial Neural Networks (ANNs) . . . . . . . . . . . . . . . 43

5.4.2.1 ANN Topology . . . . . . . . . . . . . . . . . . . . . 43

5.4.2.2 Employ ANN to Control DSTATCOM . . . . . . . . 45

5.4.3 ANNs Structure . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.4.3.1 AC Voltage Regulator Proportional Gain ANN . . . 46

5.4.3.2 AC Voltage Regulator Integral Gain ANN . . . . . . 46

5.4.3.3 DC Voltage Regulator Proportional Gain ANN . . . 48

5.4.3.4 DC Voltage Regulator Integral Gain ANN . . . . . . 48

5.4.3.5 Current Regulator Proportional Gain ANN . . . . . 49

5.4.3.6 Current Regulator Integral Gain ANN . . . . . . . . 49

CONTENTS xii

5.4.3.7 PLL Proportional Gain ANN . . . . . . . . . . . . . 51

5.4.3.8 PLL Integral Gain ANN . . . . . . . . . . . . . . . . 51

5.4.4 Designing and Programming . . . . . . . . . . . . . . . . . . . 52

5.5 Total Harmonic Distortion . . . . . . . . . . . . . . . . . . . . . . . . 53

5.5.1 Harmonics Formulation . . . . . . . . . . . . . . . . . . . . . . 53

5.5.2 Measurement of THD% In Distribution System . . . . . . . 54

6 Results and Conclusions 55

6.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 55

6.1.1 Creation of Sag and Swell Events on Distribution Bus . . . . . 55

6.1.2 Selection of Controller Gains Using PSO . . . . . . . . . . . . 58

6.1.3 Control DSTATCOM using ANNs . . . . . . . . . . . . . . . . 59

6.1.4 Effect of ANNs Controller on Performance of DSTATCOM for

Improve the Voltage Profile . . . . . . . . . . . . . . . . . . . 64

6.1.5 THD Measurement . . . . . . . . . . . . . . . . . . . . . . . . 65

6.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

6.3 Future Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

A MATLAB Codes 80

B Real feeder given by JDECO 86

C Optimum values of controller gains for swell and sag voltage events

89

List of Figures

3.1 Short duration voltage events . . . . . . . . . . . . . . . . . . . . . . 15

3.2 Quadrature method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4.1 Equivalent model of STATCOM [21] . . . . . . . . . . . . . . . . . . 21

4.2 Equivalent circuit of STATCOM [21] . . . . . . . . . . . . . . . . . . 22

4.3 Phasor diagram for the system (R not neglected) [46] . . . . . . . . . 22

4.4 Phasor diagram for the system (R neglected) [46] . . . . . . . . . . . 23

4.5 A functional model of STATCOM [21] . . . . . . . . . . . . . . . . . 24

4.6 STATCOM V-I characteristic [45] . . . . . . . . . . . . . . . . . . . 25

4.7 Configuration of power system with DSTATCOM [22] . . . . . . . . 26

4.8 DSTATCOM control strategy [22] . . . . . . . . . . . . . . . . . . . 27

4.9 Direct PI control of DSTATCOM [49] . . . . . . . . . . . . . . . . . 28

5.1 Single line diagram for radial network . . . . . . . . . . . . . . . . . 33

5.2 Matlab Simulink simulation . . . . . . . . . . . . . . . . . . . . . . . 36

5.3 MATLAB Simulink model for DSTATCOM configurations [39] . . . 37

5.4 Control system block diagram for DSTATCOM [39] . . . . . . . . . . 37

xiii

LIST OF FIGURES xiv

5.5 Control strategy for DSTATCOM [39] . . . . . . . . . . . . . . . . . 38

5.6 PSO search mechanism of PSO in multidimensional search space [44] 40

5.7 Flowchart of PSO alogorithm . . . . . . . . . . . . . . . . . . . . . . 41

5.8 Basic neural network structure [51] . . . . . . . . . . . . . . . . . . . 44

5.9 AC voltage regulator proportional gain ANN . . . . . . . . . . . . . . 47

5.10 AC voltage regulator integral Gain ANN . . . . . . . . . . . . . . . . 47

5.11 DC voltage regulator proportional gain ANN . . . . . . . . . . . . . . 48

5.12 DC voltage regulator integral gain ANN . . . . . . . . . . . . . . . . 49

5.13 Current regulator proportional gain ANN . . . . . . . . . . . . . . . . 50

5.14 Current regulator integral gain ANN . . . . . . . . . . . . . . . . . . 50

5.15 PLL proportional gain ANN . . . . . . . . . . . . . . . . . . . . . . . 51

5.16 PLL integral gain ANN . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.17 Basic flowchart for designing ANNs model . . . . . . . . . . . . . . . 53

6.1 Extreme sag event 0.7pu . . . . . . . . . . . . . . . . . . . . . . . . . 56

6.2 Extreme swell event 1.3pu . . . . . . . . . . . . . . . . . . . . . . . . 56

6.3 Extreme sag and extreme swell events . . . . . . . . . . . . . . . . . . 57

6.4 Sag events via error% . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

6.5 Swell events via error% . . . . . . . . . . . . . . . . . . . . . . . . . . 58

6.6 Training validation and test curves for neural networks (1-4) . . . . . 60

6.7 Training validation and test curves for neural networks (5-8) . . . . . 61

6.8 Estimation and fitting for neural networks (1-4) . . . . . . . . . . . . 62

6.9 Estimation and fitting for neural networks (5-8) . . . . . . . . . . . . 63

LIST OF FIGURES xv

6.10 Voltage profile improvement for extreme swell event . . . . . . . . . 66

6.11 Voltage profile improvement for extreme sag event . . . . . . . . . . 66

6.12 rms voltage profile improvement using DSTATCOM tuned by devel-

oped DSTATCOM controller using AI . . . . . . . . . . . . . . . . . 67

6.13 Phase voltage profile improvement using DSTATCOM tuned by de-

veloped DSTATCOM controller using AI . . . . . . . . . . . . . . . . 67

6.14 THD for swell events . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.15 THD for sag events . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.16 Frequency spectrum for extreme sag event . . . . . . . . . . . . . . . 69

6.17 Frequency spectrum for extreme swell event . . . . . . . . . . . . . . 69

B.1 Abu Mashaal case study given by JDECO . . . . . . . . . . . . . . . 86

B.2 Load and Voltage Profile . . . . . . . . . . . . . . . . . . . . . . . . 87

B.3 Der Abu Mshaal Data Feeder . . . . . . . . . . . . . . . . . . . . . . 88

List of Tables

3.1 PQ events classification (IEEE std 1159) [30] . . . . . . . . . . . . . . 16

5.1 Description of proposed distribution feeder . . . . . . . . . . . . . . 34

5.2 Parameters of PSO algorithm . . . . . . . . . . . . . . . . . . . . . . 42

5.3 Define networks in ANN algorithm . . . . . . . . . . . . . . . . . . . 46

5.4 Voltage distortion according to IEEE 519-1992 [39] . . . . . . . . . . 54

6.1 Optimal values of DSTATCOM controller gains during main steps of

swell events using PSO . . . . . . . . . . . . . . . . . . . . . . . . . . 57

6.2 Optimal values of DSTATCOM controller gains during main steps of

sag events using PSO . . . . . . . . . . . . . . . . . . . . . . . . . . 57

6.3 Evaluation of MSE and fitting function R for eight neural networks . 64

6.4 Voltage profile improvement using developed DSTATCOM controller 65

C.1 DSTATCOM Controller Gains For Swell Events Using PSO . . . . . 90

C.2 DSTATCOM Controller Gains For Sag Events Using PSO . . . . . . 91

xvi

List of Abbreviations

The following table lists all abbreviations and acronyms used throughout the thesis

and the page on which each is defined or first used.

Abbreviation Description

AC Alternating Current

AI Artificial Intelligent

ANNs Artificial Neural Networks

ASD Adjusable Speed Drive

BESS Battery Energy Storage

DE Differential Evaluation

DGs Distribution Generations

DSTATCOM Distribution Static Synchrounous Compensator

EESS External Energy Storage System

FACTs Flexible Alternating Current Transmission Systems

FFT Fast Fourier Transform

GA Genetic Algorithm

HS Harmony Search

IEEE Institute of Electrical and Electronics Engineers

xvii

LIST OF TABLES xviii

IWD Intelligent Water Drop

JDECO Jerusalem District Electricity Company

3LNPC Three Level Neutral Point Clamped

MI Modulation Index

NPC Neutral Point Clamped

PF Power Factor

PI Proportional Integral

PCC Point of Common Coupling

PQ Power Quality

PSO Particle Swarm Optimization

PWM Pulse Width Modulation

RES Renewable Energy Sources

RMS Root Mean Square

RMSE Root Mean Square Error

SA Simulated Annealing

SC Super Capacitor

SOA Seeker Optimization Algorithm

SPWM Sinusoidal Pulse Width Modulation

SRF Synchronous Reference Frame

THD Total Harmonics Distortion

VRB Vanactium Redox Battery

VSC Voltage Source Converter

VSI Voltage Source Inverter

VVO Volt VAR Optimization

Chapter 1

Introduction

1.1 Background

The rapid increase in electrical power demand lead emerging use of renewable energy

sources (RES) and distributed generations (DGs) in electrical grid. Emerging new

resources of energy into the grid and connecting different types of loads such as non-

linear loads create different problems appears in current, voltage and/or frequency

which causes error and malfunction of instrument and generally consider as power

quality (PQ) problems [1].

Several studies are done regarding PQ issues using the benefits of huge develop-

ments in power electronics technologies, hence the network change from traditional

network to smart network [2], [3]. Also, PQ may defined as electrical constraints that

allows an equipment function properly with high performance [4], [5]. According to

these definitions of PQ, poor PQ will affect badly on the reliability and stability of

electrical network as well as on sensitive loads connected to the end user of distri-

1

CHAPTER 1. INTRODUCTION 2

bution bus. PQ is relate to the three main factors: voltage, current and frequency,

all these factors should be remain within their standard limits [3], [4].

The problems of PQ in the network caused by different reasons such as connect-

ing nonlinear loads especially large motors in industrial applications, power elec-

tronics devices, switching capacitor banks and connected new electrical resources to

the grid [4], [5].

Indeed, these disturbances affect on the performance and functionality of both

distribution system and sensitive equipment. The smart grid (SG) technology open

a wide range for solving PQ problems [2], [3]. Family related to power electronics

technologies called flexible alternating current transmission systems (FACTs) sup-

port SG system and provide solutions for PQ challenges. One application of FACTs

family is called distribution static synchronous compensator (DSTATCOM), which

is a shunt device connected with distribution bus.

Huge development in artificial intelligence (AI), electronic control and commu-

nication capabilities contribute the development of optimizations issues of power

system aspects under SG umbrella. One of the main aspects of power system is

voltage var control. So the voltage var optimization (VVO) is the main SG appli-

cation that enhances both security and economy of power system [6].

In this work the voltage profile is improved to be within the standard acceptable

levels by using DSTATCOM with developed control strategy. The control strategy

is developed by using ANN as an AI technique, which also depends on optimum

values of PI controller obtained by PSO. The study validated and tested on the

proposed test feeder under MATLAB environment software package.


1.2 Thesis Contribution

The contribution of this thesis is developing controller for DSTATCOM based on

AI Technique ANN, which its controller gains optimally obtained by PSO.

1.3 Thesis Objectives

The main objective of this thesis has been to examine using DSTATCOM with

developed control strategy on distribution part of proposed electrical feeder in order

to improve the voltage profile. To satisfy the defined main objective the following

intermediate objectives are fixed to properly divide the work to be carried out:

1. Selecting a DSTATCOM for improving voltage profile for medium voltage

feeder.

2. Developing control strategy for the proposed DSTATCOM based on AI tech-

nique ANN, which depends on optimum values of PI controller gains using

PSO.

3. Simulating the proposed DSTATCOM on real feeder using MATLAB/Simulink

package to validate the efficient of the developed control technique.

1.4 Thesis Organization

The remaining chapters of this thesis are organized as follows:

Chapter 2 Extensive literature reviews are discussed about PQ issues and their

effects on distribution system performance. A considerable amount of research are


reviewed regarding the voltage profile improvement using DSTATCOM and their

different control strategies used for this purpose. Also, reviewed researches deal

with increase the performance of DSTATCOM using RES. Finally, the reviewed the

AI techniques used for tuning PI controller of DSTATCOM and validate the research

gap to do this research.

Chapter 3 provides an overview of PQ problems in distribution system their

causes and classifications of events and discuss the solutions and evaluation of PQ

issues in distribution system.

Chapter4 presents DSTATCOM operation, control strategy, mathematical

model, DSTATCOM design equations.

Chapter 5 presents the case study for this research. Also, discussed the AI

technique, which is ANN as well as the optimization technique which is PSO. In

addition, MATLAB model is simulated according to the studied feeder (11kV/50Hz)

for Abu Mashaal zone in Rammallah city. MATLAB design depends on given data

for real electrical feeder. Then algorithms for ANN and PSO are discussed in details

for a purpose to use them in developed controller.

Chapter 6 discuss the simulations results for developed controller of DSTAT-

COM and the improvement of voltage profile for extreme sag and extreme swell

voltage events. Then proved the validation and the efficient of developed controller.

Also, THD is evaluated for improved voltage signal. All in all, conclusions inferred

during this work. In addition, ideas and recommendations drawn to be done as

future work.

Chapter 2

Literature Review

2.1 PQ Problems in Power System

Several research studies are done regarding improvement PQ issues in electrical

power system [1]- [4], [7]. More recent attention has focused on provision regarding

enhancement the reliability and stability of electrical grid by means of improving

the PQ issues. These research discussed problems of PQ such as voltage profile

regulation, unbalance conditions causes by faults and load changing, system losses

due to poor power factor (PF) and total harmonic distortion (THD). DSTATCOM

and other FACTs devices with different control strategies are used to solve PQ

challenges [8], [9].

5

CHAPTER 2. LITERATURE REVIEW 6

2.2 Voltage Profile Improvement Using DSTAT-

COM

Several research studies validated the significant effect of DSTATCOM on PQ issues.

In [10] the research proved the efficient performance of DSTATCOM for improving

the voltage profile dynamically with high penetration level of wind turbine con-

necting on distribution bus. DSTATCOM is useful and efficient for reactive power

compensation on distribution bus since it tested for PF correction, voltage regulation

and load balancing especially where the DG is vital as alternative source of energy

in applications such as ships, aircrafts and in the remote isolated power generation

systems [11].

Research study [12] discussed the renewable DGs as a solution for voltage dip

on distribution system which contain sensitive equipments. The study did not rec-

ommended increase the high perpetration levels of DGs in the power system due

to that will increase control and coordinations complexity. The study suggested

using AI and control methodology for fast acting mitigation of voltage dip. Other

study [13] supported study in [12] by validated DSTATCOM efficiently for mitiga-

tion of voltage sag as well as correct PF by compensation enough reactive power.

Also in [14] a group of researchers examined using STATCOM on transmission sys-

tem (132kV/50Hz) for compensation the voltage swell and the results validated

STATCOM as efficient FACTs device for improving voltage profile on transmission

system.

On the other hand, in [15] many PQ issues are mitigated including voltage

regulation with unity power factor, harmonics distortion and unbalance voltage.


The study is implemented on the proposed network. PQ problems with different

types of load examined were mitigated and DSTATCOM proved to use as a solution

for improving power system stability and reliability.

2.3 Control Strategies of DSTATCOM

A considerable amount of literature has been published on DSTATCOM control

strategies. The PWM connected to distribution bus by LCL filter is implemented

using digital signal processing (DSP) controller in order to eliminate negative se-

quence for unbalance voltage and so improve the voltage profile during fault condi-

tions. The results proved the vital of using LCL filter with DSTATCOM in order

to mitigate the PQ problems [8], [16].

Different control strategies used in DSTATCOM are sensitive to non ideal supply

conditions and load variation which needed careful design tuning controller for each

specific insulation. Also, the cost of the system increase due to use special devices

such as LCL filter [8], [16]- [17].

Other control strategies of DSTATCOM in literature needed complex control fa-

cilities and controllers, also needed additional supplementary components such as

active filter. In addition, the controller may need additional delay time due the

conversion from analogue to digital [17]- [19].

On the other side new proposed control strategies for DSTATCOM are imple-

mented such as adaline based algorithm and direct method which depends mainly

on position and magnitude control of output voltage. Although these methods are

validated as efficient methods for controlling DSTATCOM, they are complex and

expensive comparing with traditional methods [9], [18].


Different mathematical theories and physical models are used in analysis and

implementing different control strategies of DSTATCOM. The cascaded technique is

used and the voltage source inverter is linearized using average switch model(ASM)

[8]. Instantaneous power theory, rotating reference frame(dq0), stationary reference

frame(α β0), instantaneous reactive power (IRP) and Park’s transformation are used

to simplify the dynamic system equations [8], [9], [16]- [21]. The stationary reference

frame transform balance three phase system into two phase (clark transformation)

this reduce one phase variable implies the same reduction in the number of controllers

required, which will reduce the complexity of the control system. On the other hand,

dq0 provides a steady state control variable which is more simplicity and efficient for

controller design [22]. The Phase Locked Loop(PLL) is used to provide a reference

phase signal synchronize with the ac system [20]. Proportional Integral (PI) used

to regulate and control firing angle provided to PWM, which generated switching

pulses to IGBTs in VSI [18].

2.4 Merging RES with DSTATCOM in SG

Recently a SG technologies play an important role for improving reliability of elec-

trical power system. SG contribute the improving of PQ issues with corporation of

power electronics technologies and RESS. Few papers discussed the integration be-

tween RESS and DSTATCOM to improve PQ issues in the network [7], [22]. Signifi-

cant research has been concerned with FACTs devices and smart control compatible

with renewable energy resources. Existing RES and DGs in power system create

different challenges especially when these resources with high penetration levels. To

deal with these challenges a SG technology is vital to manage the power flow gener-


ated by grid. Power generated from RES and the compensation power generated or

absorbed by FACTs devices. Using SG technology leads to increase the reliability

and stability of power system regardless of high perpetration level of RES connected

to distribution bus [2], [6], [23]- [24].

The dynamic performance of DSTATCOM is studied by coupling it with exter-

nal energy storage system (EESS) in order to enhance PQ at distribution system.

Multilevel control techniques are used in VSI. Battery was used as energy storage

system (BESS) and added to dc bus of the inverter. This is efficient to reduce invest-

ment cost by avoiding extra coupling interface. The VSI designed using multilevel

topology which is neutral point clamped (NPC) provides conventional sinusoidal

output voltage without increasing switching frequency, consequently lower losses

and reducing size, but this topology needed to maintain a constant voltage at half

dc bus voltage to prevent large distortion [20].

PHD thesis was done at ESTIA Institute of technology about using microgrid hy-

brid energy storage integration and control using a three level neutral point clamped

(3LNPC) converter. The author studied the effect of microgrid in distribution sys-

tem on power quality under high penetration level. In order to maintain stable

power system and high power quality factors in distribution system. Special storage

devices with specific requirements are used. The thesis focused on designing a power

converter system to satisfy the defined requirement of hybrid energy storage system

in order to improve the power quality under high penetration level. Two storage

devices are used , super capacitor (SC) bank and vanactium redox battery (VRB)

were examined to get suitable storage system. This hybrid system is controlled

by using 3LNPC, which select the current reference of hybrid storage system and


controlled it by typical two loop PI controller. The power and energy management

was studied to reduce the power losses and harmonic distortion. The control strat-

egy is applied on dc side of 3LNPC converter and hence there is no effect on ac

side. The author concluded that using hybrid energy storage system using SC and

VRB is essential when there is high penetration level from microgrid connecting to

distribution system [18].

2.5 Tuning DSTATCOM Controller Using AI

Techniques and Optimization Methods

Huge development in AI, electronic control and communication capabilities con-

tribute the development of optimizations issues of power system aspects under SG

umbrella [25]- [42]. On other side, several research done regarding optimization

methods for different issues in power system. These research include a comparative

of meta-heuristic approach such as GA, PSO, differential evaluation(DE), harmony

search(HS) and seeker optimization algorithm(SOA) [43]- [45].

In [26] nonlinear controller of DSTATCOM which connected to distribution sys-

tem with DGs is tunned by using the controller parameters of PID controller of

DSTATCOM is tunned using hybrid combination algorithm consists of simulated

annealing(SA) and intelligent water drop(IWD). SA is motivated by analogy to an-

nealing in solids [27] and IWD is a nature inspired algorithm that tries to mimic

the behavior of swarm of natural water drops which creates complex paths which is

optimum [28]. Researchers in [41] used SA as an optimization technique to identify

controller parameters of STATCOM and their results compared with other opti-


mization techniques, which are PSO and genetic algorithm (GA) under the same

disturbance conditions. SA proved its efficiency compared with other methods for

voltage profile improvement. This hybrid tuning methods leads to better regula-

tion of direct voltage for capacitor in DSTATCOM which improve the maximum

overshoot, rising time and steady state errors values. Researchers in [29] suggested

using PSO based tuning algorithm for tuning DSTATCOM controllers. They proved

experimentally the efficient of PSO algorithm for tuning PI gains of controllers and

compared the results with results obtained from using Zieglar Niclolos methods.

They supported the experimental results with simulation results which done under

MATLAB/Simulink environment. Researchers concluded that using suggested PSO

tuning is far superior than conventional Zieglar Niclolos tuning method.

2.6 Research Gap

From above literature therefore a considerable amount of research about using

DSTATCOM as an application of FACTs for PQ improvement in power system.

The research divided generally in many sectors such as benefits of using DSTAT-

COM for voltage profile, decreasing THD, PF correction and solving unbalance

conditions. On other side, DSTATCOM controller discussed in details to enhance-

ment the dynamic performance of its operations. Regarding to the control strategies

tuning DSTATCOM controller gains has a great challenge. Several algorithm are

used for optimally tunning PI controller of DSTATCOM, and the results are com-

pared between some of these methods. Although different algorithms suggested as

optimally tuning technique for DSTATCOM controller, the results are different with

the objective function of optimization as well as dependent on the data given about


the distribution system which is in some cases needed real time monitoring and

needed special control strategies as well as additional equipments, this lead to the

importance of using AI techniques for controlling DSTATCOM.

From the literature review, therefore great challenges regarding using AI tech-

niques for tuning PI controller of DSTATCOM. This work will suggest using ANNs

as a technique for tuning controller of DSTATCOM, which its controller constants

is selected optimally using PSO.

Chapter 3

Power Quality Factors

3.1 PQ Concept

PQ of distribution network is relate of powering and grounding sensitive equipment.

Also PQ may defined as electrical constraints that allows an equipment to function

properly with high performance. According to these definitions of PQ, poor PQ will

affect badly on the reliability and stability of electrical network as well as on sensitive

loads connected to grid. PQ is relate to the three main factors: voltage, current and

frequency, all these factors should be remain within there standard limits [2, 3].

3.2 Importance of PQ Factors in Power System

The following major reasons may lead the importance of PQ factors in power system

[30]:

• Huge development in power electronics technologies and microprocessor based

13

CHAPTER 3. POWER QUALITY FACTORS 14

control in newer generation are sensitive to PQ factors variations.

• It essential to reduce losses, improving power factor and decreasing THD espe-

cially with industrial applications such as adjustable speed motor drives(ASD).

• SG technologies may depend on the integration and interconnection between

the process of whole components in power system, this means any failure

in any component may has great consequences on the normal operation and

functionality of power system.

• Interruption of power system due the PQ issues will affects badly on the whole

international economics. Accordingly, the above the electric utilities and users

should concern about improve the PQ factors as well as electrical equipment

manufacturers.

3.3 Voltage Events of PQ

3.3.1 Classifications Voltage Events

Voltage events of PQ are divided into short duration and long duration [30]:

1. Voltage sag(dip) : according to the IEEE Std.1159 define as decrease in RMS

voltage to range between 10% to 90% for duration from 0.5 cycle to 1 min.

Voltage sag may appear when a faults happened, turning on heavy loads or

turning off large capacitor. Sag event have undesirable effects especially on

sensitive electrical devices such as ASD.


2. Voltage swell : according to IEEE Std.1159 define as an increase in RMS

voltage above 110% to 180% of nominal voltage for duration of 0.5 cycle to 1

min. Voltage swell is occurred in case of unsymmetrical single line to ground

fault, switching off a large load or when energizing a large capacitor bank. The

effects of this event are memory loss, data errors, bright lights, false tripping

and equipment shutdown [30].

3. Voltage interruption : according to IEEE Std.1159 occurs when the supply

voltage decreases to less than 10% of nominal RMS voltage for time period

not exceeding 1 min. Different causes for interruption in network such as

utility faults, circuit breaker tripping and components failure. The effect of

interruption are loss of data when false shutdown happened damage the elec-

tronics devices. Figure 3.1 shows the three short duration events and RMS

voltage waveforms [30].

Figure 3.1: Short duration voltage events [30]

Table 3.1 [30] summarizes the PQ classification regarding their time range according

IEEE std.1159


Table 3.1: PQ events classification (IEEE std 1159) [30]

Voltage Event Magnitude Duration

Sag 90%-10% ≤ 1 min

Under Voltage 90%-10% > 1 min

Interruption <10% ≤ 1 min

Sustain Interruption <10% > 1 min

Swell >110% ≤ 1 min

Over Voltage >110% > 1 min

3.3.2 Measurement of Duration and Amplitude of Voltage

Event

According to IEEE Std.1159 and IEC (6100-4-30) the events quantified by measuring

duration and amplitude. Two methods used to calculate the RMS value

1. Conventional RMS Calculation Method. The RMS value can be calculated if

the waveform is sampled given by Eq. (3.1) [31]

V rms(i) =

√√√√ 1

N

N∑i=1

(V i)2 (3.1)

where :

N : is the number of samples per cycle .

V : is the instantaneous sample voltage .

k : is the instant when the rms voltage is estimated.

2. The Quadrature Method [31] This method also used to calculate RMS voltage


based on the sampled time domain , but it uses only two samples per a half

cycle with 90 degree distance between them Figure 3.2 [31]

Figure 3.2: Quadrature method [31]

The following equations derive the value of RMS voltage using quadrature

method

V (t) = Vpsin(wt) (3.2)

S1 = Vpsin(θ) (3.3)

S2 = Vpsin(θ +π

2) (3.4)

S2 = Vpcos(θ) (3.5)


S21 + S2

2 = V 2p (sin2(θ) + cos2(θ)) (3.6)

√S21 + S2

2 =√V 2p (3.7)

√S21 + S2

2 = Vp (3.8)

Vrms =Vp√

2=

√S21 + S2

2√2

(3.9)

where :

V(t) : instantaneous voltage .

Vp : peak voltage .

S1 : sample number 1 .

S2: sample number 2 .

Vrms : root mean square voltage .

3.4 PQ Problems Evaluation

In order to deal with PQ problems. Generally, several steps may consider as follows

[32]:

1. Identification of problem category. Monitoring the system voltage, current

and frequency on the define time period lead to classify the PQ problems such


as voltage regulation, voltage unbalance, voltage sag, voltage swell, voltage

interruption, flicker, transients, harmonic distortion and other problems.

2. Problem characterization the data measurements of the system give indications

about the problems characteristics, causes and equipments impacts.

3. Identification range of solutions diagnose the problems according to data col-

lection leads to the range of solutions. These solutions may in utility trans-

mission system, utility distribution system, end use customer interface, end

use customer system or equipment design/specifications.

4. Evaluate range of solutions. The modeling and analysis procedures done to

select the suitable solutions or the alternatives.

5. Finally find the optimum solution which is the best solutions and also the

most economic one between the range of solutions and alternatives.

Chapter 4

DSTATCOM Configuration and

Design

4.1 Principle and Operation of DSTATCOM

4.1.1 Equivalent Model of DSTATCOM

DSTATCOM consists of three phase bus, shunt transformer, VSC and DC capaci-

tor. Figure.4.1 shows the equivalent model of DSTATCOM [21].

20

CHAPTER 4. DSTATCOM CONFIGURATION AND DESIGN 21

Figure 4.1: Equivalent model of STATCOM [21]

Where:

Rs and Ls: DSTATCOM transformer resistance and inductance respectively .

Eabc: are converter AC side phase voltages.

Vabc: are the ac system side phase voltages.

iabc: are phase currents.

Id: is capacitor current.

Cdc: is dc capacitor value.

The main electronic part of DSTATCOM is VSC, the function of this part is to

convert dc input voltage to ac output voltage. There two types of VSC used:

1. Square wave inverters using Gate Turn-off Thyristor four three level inverters

are utilized to make 48-step voltage waveform, beside that special inter con-

nection transformers are used to eliminate harmonics from the output voltage.

2. PWM, this inverter use isolated gate bipolar transistor (IGBT). This type

used PWM technique to convert dc voltage to ac sinusoidal waveform.

The function of coupling shunt transformer is to connect the converter to high


voltage power system. From the above model the equivalent circuit of STATCOM

simplified as shown in Figure4.2 [21].

Figure 4.2: Equivalent circuit of STATCOM [21]

Where:

Vac: ac voltage of the system

Vout: output voltage of the converter

Iac: ac current flows through the reactance between the converter and ac system

The phasor digram for the system when resistance of the transformer not ne-

glected shown in Figure4.3 [46].

Figure 4.3: Phasor diagram for the system (R not neglected) [46]

The resistance of the transformer is very small, it may be neglected. So that,

there is no dissipated power in ideal DSTATCOM. The phasor diagram for the


system shown in Figure4.4 [46].

Figure 4.4: Phasor diagram for the system (R neglected) [46]

4.1.2 Principle of Operation

To study the operation of STATCOM, a functional model of the device drawn as

shown in Figure4.5 [21].

The operation of STATCOM is summarized by three modes:

• Mode1: STATCOM supplies reactive power to AC system. If the amplitude of

Vout is increased above that of utility bus voltage Vac, the current flows through

the reactance from the converter to AC system and the converter generates

capacitive reactive power for the AC system. The phasor diagram shown in

Figure4.4 [46] the current lag Vx but this current leads Vac.

• Mode2: STATCOM absorbs reactive power from AC system. If the amplitude

of Vout is decreased below the utility bus voltage, the current flows from the

AC system to the converter which absorbs inductive reactive power from the


Figure 4.5: A functional model of STATCOM [21]

AC system. From the phasor diagram shown in Figure4.4 , the current lag Vx

also the current lag Vac.

• Mode3: STATCOM is in floating state. If the Vout equal the AC system voltage

there is no exchange reactive power.

4.1.3 STATCOM V-I Characteristic

When the STATCOM operate as voltage regulation on distribution system. The

V-I curve describes its operation as shown in Figure4.6 [45].

from the V-I characteristic the STATCOM supply both capacitive and inductive


Figure 4.6: STATCOM V-I characteristic [45]

compensation and it is able to independently control its output current over maxi-

mum capacitive and maximum inductive as long as the reactive current stays within

(-Imax in capacitive mode and Imax in inductive mode). The slope regulation is

define as difference between the reference voltage and voltage bus of ac system.

DSTATCOM has a nonlinear structure, so that in order to study the operation

and it control strategy a simplified mathematical formula is needed. Figure4.7 shows

the configuration parameters of power system with DSTATCOM [22]. DSTATCOM

operation based on the fact that the real and reactive power can be adjusted by

voltage magnitude of the inverter Vc and the angle difference between the bus and

the inverter output (α). The Eq. (4.1) and Eq. (4.2) are describe the active and

reactive power compensation respectively [29]:

P =VpccVcsinα

X(4.1)

Q =Vpcc(Vpcc − Vccosα)

X(4.2)


Figure 4.7: Configuration of power system with DSTATCOM [22]

Where:

P: Active Power.

Q: Reactive Power.

Vc: Inverter Voltage.

Vpcc: Voltage at the point of common coupling.

α: Angle of Vpcc with respect to Vc.

X: Reactance of the branch and the transformer.

DSTATCOM connected to the grid through coupling transformer at PCC exchanges

reactive power with the grid. If the magnitude of the DSTATCOM voltage Vc is

greater than the grid voltage Vpcc (Vc>Vpcc), then DSTATCOM supplies reactive

power to the grid and hence DSTATCOM is operating in the capacitive mode. If

the grid voltage Vpcc is greater than the DSTATCOM voltage (Vpcc>Vc) the DSTAT-

COM absorbs reactive power from the grid and then DSTATCOM is operating in

inductive mode. If the grid voltage (Vpcc) and DSTATCOM voltage (Vc) are the same

magnitude (Vpcc=Vc) there is no exchange of reactive power between the grid and


DSTATCOM. On the other hand, active real power can flow through the DSTAT-

COM in either direction from DC link capacitor to PCC and from PCC toward the

DC link capacitor.

4.2 DSTATCOM Control Strategy

The function of DSTATCOM under justified condition is to inject the required

reactive current to the load. DSTATCOM inject these current by using PI controller.

The control strategy scheme is shown in Figure4.8 [22]. The output voltages for VSI

Figure 4.8: DSTATCOM control strategy [22]

are controlled by PWM, which is related to modulation index (MI) and phase angel

α. The required control variables are generated by adjustable tuned PI controller

in DSTATCOM scheme.

The output voltage of DSTATCOM is controlled in such away that the phase

angle between the inverter voltage and line voltage is dynamically adjusted so that

the DSTATCOM generates or absorbs the desired reactive power at PCC [47]. The


goal of such control is to maintain constant voltage magnitude at PCC under system

disturbances [49]. Figure4.9 shows the controller input is an error signal which is

Figure 4.9: Direct PI control of DSTATCOM [49]

the difference between the RMS reference voltage(1pu) and the RMS value of the

terminal voltage measured. The PI controller process the error signal and generate

the required drive angle to minimize the error to zero.

4.3 System Mathematical Model

A according to the dynamic equations of the system, the controller of DSTATCOM

is designed. Assuming that the system is considered as linear system [22].

4.3.1 Load Model Mathematical Equations

According to system configuration shown in Figure4.7, the three phase load voltages

coordinates may written as in Eq. (4.3) [22] :

vla

vlb

vlc

= Rl

ila

ilb

ilc

+ Lld

dt

ila

ilb

ilc

(4.3)


The load is assumed to be balanced, the park transformation is used to convert three

phase system in to two phase dq system.

The park transformation [6] is given as in Eq. (4.4)

T =2

3

cosθ cos(θ − 2π

3) cos(θ + 2π

3)

−sinθ −sin(θ − 2π3

) −sin(θ + 2π3

)

12

12

12

(4.4)

Then the Eq. (4.3) can transformed to rotating reference frame as in Eq. (4.5) [22]

: vldvlq

= Rl

ildilq

+ Lld

dt

ildilq

+ Ll

0 −w

w 0

ildilq

(4.5)

where:

θ = tan−1(vlq/vld)

4.3.2 VSI AC-Side Model

The source voltage Vs and inverter output voltage may be written as given in Eq.

(4.6) and Eq. (4.7) respectively [22].

Vsa

Vsb

Vsc

=a1Ll

isa

isb

isc

+b1Ll

d

dt

iea

ieb

iec

+c1L1

iea

ieb

iec

+d1Ll

ea

eb

ec

(4.6)

CHAPTER 4. DSTATCOM CONFIGURATION AND DESIGN 30ea

eb

ec

=a2Ll

isa

isb

isc

+b2Ll

d

dt

isa

isb

isc

+c2L1

iea

ieb

iec

+d2Ll

Vsa

Vsb

Vsc

(4.7)

Where:

a1=(RsLl −RlLs)

b1=-(LsLf + LsLl + LlLf )

c1=-(RfLs +RlLs +RfLl)

d1=(Ls + Ll)

a2=-(RsLf +RlLf +RsLl)

b2=-(LsLf + LfLl + LlLs)

c2=(RfLl −RlLf )

d2=(Lf + Ll)

4.3.3 VSI DC-Side Model Equations

The power balance equation for VSI [6] is expressed as Eq.(4.8) [22]

vdcidc =3

2(edied + eqieq) (4.8)

vdc can be derived from a current balancing formula as in Eq.(4.9) [22] :

d

dtvdc = −

(vdc

RdcCdc+idcCdc

)(4.9)


The current idc can be expressed in Eq.(4.10) [22]:

idc =3

2vdc(edied + eqieq) (4.10)

4.4 DSTATCOM Design

According to the mathematical model equations of the system the selection of

DSTATCOM components should be calculated [34]

4.4.1 Design DC Bus Voltage

Sufficient enough DC link voltage is vital to obtain current compensation which can

be calculated as in Eq.(4.11) [34]:

Vdcmin >√

2VL−L(r.m.s) =√

2√

3VL−N(r.m.s) (4.11)

4.4.2 Design DC Bus Capacitor

Referring to the principle of energy conservation, the Cdc may calculated in Eq.

(4.12) [34]

1

2Cdc[V

2dc − V 2

dc1] = 3KaV It (4.12)

Where:

Vdc: is the reference DC bus Voltage.

Vdc1: is the minimum level of the DC bus voltage.

a: is the over loading factor.


V: is the phase voltage.

I: is the phase current.

t: is the time for which the DC bus voltage is to be recovered.

k: factor for variation of energy during dynamics.

4.4.3 Design AC Interfacing Inductor(Lf)

The selection of suitable filter depends on the ripple current(icr−pp), switching

frequency(fs) and DC bus voltage (Vdc) is given in Eq. (4.13) [34]

Lf =

√3mVdc

12afs ∗ icr pp(4.13)

Where:

m: is modulation index.

4.4.4 Design Passive High Pass Ripple Filters

This filter is need to deal with high frequency noise from the voltage at PCC and it

may calculated as given in Eq.(4.14) [34]:

RrCr =Ts10

(4.14)

Chapter 5

Methodology

5.1 Case Study

The work is applied on real network with data given by Jerusalem District Elec-

tricity Company (JDECO). The network is for Abu Mashaal zone in West Bank

(appendix B). The voltage on this network drops below the standard level (11kV)

on distribution bus due the industrial nonlinear loads. The single line diagram of the

given radial network is shown in Figure 5.1 Description of the system configuration

Figure 5.1: Single line diagram for radial network

33

CHAPTER 5. METHODOLOGY 34

is summarized in table. 5.1,the details data is given in (appendix B).

Table 5.1: Description of proposed distribution feeder

Parameter Value3phase ac source Vs(rms)= 11kV, f=50HzLine impedance Ls =3.851mH,Rs =0.121ΩDistribution transformer (3kVA) 1. 33 kV/11kV

2. 11kV/0.4kV

5.2 Calculation of DSTATCOM Configurations

Configurations of DSTATCOM are calculated using mathematical equations design

given in Eq.(4.11) into Eq.(4.14).

The rated load connected on the end user of distribution system is 1MVA, according

to this the rating current may calculated as follows:

Rated RMS current =kV A ∗ 103

√3 ∗ 400

=1000 ∗ 103

√3 ∗ 400

= 1443.3A (5.1)

The Peak value of line current may calculate:

IPeak = 1443.3 ∗√

2 = 2041.2A (5.2)

The DC link Bus Voltage calculate:

Vdc =√

2 ∗ 11kV = 15556.3V (5.3)


The DC bus capacitor value may estimated as:

Cdc =[3 ∗ 0.1 ∗ 1.5 ∗ 6350.8 ∗ 2041.2 ∗ 20 ∗ 10−3] ∗ 2

(16000)2 − (15556.3)2= 16.665mF (5.4)

The AC interfacing inductor (Lf) calculate as:

Lf =

√3(1) ∗ 16000

12 ∗ 1.2 ∗ 200 ∗ 103 ∗ (0.15 ∗ 2041.2)= 31.42µH (5.5)

The RrCr high pass filter design as:

RrCr =Ts10

=5 ∗ 10−6

10(5.6)

Let Cr=1.8*10−3 F , ⇒Rr= 0.0027Ω

5.3 MATLAB Model Simulation

5.3.1 MATLAB Simulink Simulation

Figure 5.2 shows the over all power system for given case study with DSTATCOM

connecting on 11kV distribution bus

5.3.2 MATLAB Simlink For DSTATCOM

The DSTATCOM configuration is simulated as shown in Figure 5.3 [39]. There are

two IGBT bridges in DSTATCOM configurations this may reduce the harmonics.


Figure 5.2: Matlab Simulink simulation

5.3.2.1 DSTATCOM Controller Simulation

DSTATCOM control strategy mainly based on PI controller used to regulate the

error and so minimize the voltage derivation between the reference value and the

actual measured value. The controller performance depending upon the character-

istics of the plant and the values of Kp and Ki parameters used in PI controller.

Figure5.4 [39] shows the close loop control system contain Function of PI controller

Gc(s) and objective function G(s). In order to derive the transformer function of

control system which given in Eq. (5.7) [39] :

TransferFunction =Gc(s) +Gs

1 +Gc(s)Gs

(5.7)


Figure 5.3: MATLAB Simulink model for DSTATCOM configurations [39]

Figure 5.4: Control system block diagram for DSTATCOM [39]

The PLL is used as a control strategy for DSTATCOM. The control strategy is

shown in Figure5.5 [39]. The controller of DSTATCOM consists of PLL, measure-

ment systems, current regulation loop, voltage regulation loop and a dc link voltage

regulation.

The PLL is synchronize to the fundamental of the transformer primary voltage

to provide synchronous reference trigonometric functions required for the abc-dq

transformation. The inner current regulation loop consist of two PI controllers that


Figure 5.5: Control strategy for DSTATCOM [39]

control the dq components for currents. The controllers outputs are the voltage

direct axis and quadrature axis Vd and Vq that the PWM has to generate. The

inverse park transformation is used to convert Vd and Vq back to Vabc, which are used

to synthesize the PWM voltage. The distribution bus voltage is regulated by the PI

controller which produces Iq reference for current controller. The Id reference comes

from dc link voltage regulator which maintain the dc link voltage constant [34].

5.4 Artificial Intelligence and Optimization Tech-

niques

PSO is used as a tool for select the optimum values for tuning PI controllers of

DSTATCOM. Then ANNs trained according to the PSO results. The ANNs used

to control DSTATCOM real time according to disturbance on distribution bus.


5.4.1 Particle Swarm Optimization

5.4.1.1 PSO Topology

PSO is a heuristic search optimization method. PSO inspired by social and

cooperative behaviors. The technique is related to the main parts. One is guided by

personal behavior (Pbest) and the other part guided by social experience (Gbest).

The position of particle in search space is updated depends on these two parts.

The behaviors of particles are accelerated by two factors c1 and c2 and random

number r1 and r2 numbers generated between [0, 1] whereas the present movement

is multiplied by an inertia factor ω varying between [ωmin, ωmax]. The initial

population (swarm) of size N and dimension D is denoted as X = [X1, X2, ..., XN ]T

, where ’T’ denotes the transpose operator. Each individual (particle)Xi (i =

1,2,...,N) is given as Xi=[Xi,1 , Xi,2 , ..., Xi, D]. Also, the initial velocity of the

population is denoted as V = [V1, V2, ..., VN ]T .

Thus, the velocity of each particle Xi (i = 1,2,...,N) is given as

Vi=[Vi,1 , Vi,2 , ..., Vi, D]. The index i varies from 1 to N whereas the index j

varies from 1 to D. The detailed algorithms of various methods are described below

for completeness [43]- [44].

V k+1i,j = w × V k

i,j + c1 × r1 × (Pbestki,j −Xki,j) + c2 × r2 × (Gbestkj −Xk

i,j) (5.8)

Xk+1i,j = Xk

i,j + V k+1i,j (5.9)

In Eq. (5.8) Pbestki,j represent personal best jth component of ith individual,


whereas Gbestkj represents jth component of the best individual of population up to

iteration k.

Figure5.6 [44] shows the search mechanism of PSO in multidimensional search space.

Figure 5.6: PSO search mechanism of PSO in multidimensional search space [44]

A detailed flowchart of PSO considering the above steps is shown in Figure5.7.

5.4.1.2 Tuning PI Controller Gains of DSTATCOM Using PSO

PSO algorithm is applied to find the optimum controller gains of DSTACOM (Vac

regulator, Vdc regulator, current regulator and PLL regulator). The algorithm pa-

rameters are selected as given in table. 5.2. The objective function is minimization

of root mean square error (RMSE) between the actual measured voltage and refer-

ence voltage (1pu). The RMSE and the objective function are as given in Eq. (5.10)


Figure 5.7: Flowchart of PSO alogorithm

and Eq. (5.11) respectively [48]:

RMSE =

√√√√ 1

n

n∑i=1

(Vi − Vref )2 (5.10)


Table 5.2: Parameters of PSO algorithm

Parameter SizePopulation size N=50Accelerated factors C1=1,C2=1.5Inertia w=1.2Number of iteration i=100Upper limits for controller gains [10 3500 10 20 5 1500 1500 3000]Lower limits for controller gains [0.01 10 0.0001 0.01 1 100 10 10]

J = min[RMSE] (5.11)

Where:

J: fitness function.

RMSE: Root mean square error.

Vi:actual voltage.

Vref : reference voltage(1pu).

n: number of samples.

Matlab code is written for the objective function and the PSO algorithm is running

for voltage profile improvement(appendix A). Sag and swell events are created in

order to obtain controller gains for deviation of voltage from 1pu. The simulation is

running for every deviation of (0.05pu) from extreme swell to extreme sag (appendix

C). At every case of sag and swell event the percentage error is calculated as well

the THD%.


5.4.2 Artificial Neural Networks (ANNs)

In this part an ANNs are used to control DSTATCOM real time according to dis-

tribution bus voltage level.

5.4.2.1 ANN Topology

ANN is a common statistical learning algorithm in data mining. The main idea of

ANN was inspired by the biological neural network of the human brain. The human

brain is composed of a vast number of neurons that interconnected in an incredibly

complex structure. ANN has a similar structure to the human brain; that is, it

predicts events by learning the information and stores it within the neurons by

interchanging connections and weights [50].

For many complex issues, which are hard to solve by traditional mathematical

and computation methods, ANN might be a suitable solution. As such, this research

uses a well used type of neural network model, which is known as feed forward for

tuning DSTATCOM controller gains. Figure 5.8 shows the basic architecture of a

feed forward of ANN [51]. The ANN has three layers architecture: input nodes,

hidden nodes and output nodes, each node (neurons) in one layer is connected to

each node in the next layer. The number of nodes in the input layer is the same as the

input features in the training dataset and the number of nodes in the output layer is

the number of output. The number of hidden nodes in the middle layer is normally

between those two numbers. The number of hidden nodes is very important since

both over fitting and under fitting will effect on the training results.

The procedure of training the ANN is depending on the varying the weights Wij

and the biases Bj [36]. The training criterion is taken as the mean square error of


Figure 5.8: Basic neural network structure [51]

ANN output with value of (0.0001) and error function is defined as given in Eq.

(5.12) [36]

J =N∑i=1

e(i)2 (5.12)

Where N is the number of output neurons and e(i) is instantaneous error between

the actual and estimated value of the output. The hidden layer has no target values,

so a procedure of optimize the Wij is continued until minimization the error [37].

The jth neuron in the hidden layer is assigned the value yj, given in terms of the

input values xi by [36]

yj = tanh(∑

Wijxi+ bj) (5.13)

The hyperbolic tangent function is used as activation function. Other functions

could be used for activation function, which was designed originally to simulate


the neuron upon receiving input signals from its neighbors. If the hidden layer

increases, then Eq. (5.13) will be used to calculate the values of the next layer

of hidden neurons from the current layer of neurons. The output neurons zk may

calculate by a linear combination of the neurons in layer just before the output layer

as given in Eq. (5.14) [36]

zk =∑j

Wijyj + bk (5.14)

5.4.2.2 Employ ANN to Control DSTATCOM

ANN is proposed as AI method used for controlling DSTATCOM. A PI controller

calculates an error value as the difference between a measured process and desirable

set point. The controller attempts to minimize the error by adjusting the process

through use of a manipulated variables. The PI controller algorithm involves two

separate constant parameters; P depends on the present error and I depends on the

accumulation of past errors. The weighted sum of PI components actions is used to

adjust the process via control element to control various system [38].

The environment of ANN controller has made them a wide area of interest

among researches in extensive field, due ANN can proficiently learn the unknown

continuously varying environment and act accordingly. The results from PSO for

wide range of sag and swell events (0.7-1.3) processed and nearly 70% of these

data(appendix C) considered as a data base for ANNs training process. Then the

results of ANNs provides real time to control DSTATCOM.

The learning process of ANN is developed in MATLAB, aided by the toolbox

neural network(appendix A). For this study eight networks are defined in MATLAB

code in order to applied the ANNs algorithm for learning and find the best optimal


value of controllers gains according to disturbance voltage level, the networks is

define as shown in table. 5.3:

Table 5.3: Define networks in ANN algorithm

ANNs Controller Gainsnet1 Vac (kp)net2 Vac (ki)net3 Vdc (kp)net4 Vdc (ki)net5 i(kp)net6 i(ki)net7 PLL(kp)net8 PLL(ki)

5.4.3 ANNs Structure

5.4.3.1 AC Voltage Regulator Proportional Gain ANN

This network has been designed to select the value of Vac (kp). It consists of an

input layer of 50 neurons, a hidden layer of 60 neurons and an output layer of a

single neurons. Figure 5.9 shows the structure of this network.

5.4.3.2 AC Voltage Regulator Integral Gain ANN

This network has been designed to select the value of Vac (ki). It consists of an input

layer of 40 neurons, a hidden layer of 50 neurons and an output layer of a single

neurons. Figure 5.10 shows the structure of this network.


Figure 5.9: AC voltage regulator proportional gain ANN

Figure 5.10: AC voltage regulator integral Gain ANN


5.4.3.3 DC Voltage Regulator Proportional Gain ANN

This network has been designed to select the value of Vdc (kp). It consists of an



Figure 5.11: DC voltage regulator proportional gain ANN

5.4.3.4 DC Voltage Regulator Integral Gain ANN

This network has been designed to select the value of Vdc (ki). It consists of an input




Figure 5.12: DC voltage regulator integral gain ANN

5.4.3.5 Current Regulator Proportional Gain ANN

This network has been designed to select the value of i(kp). It consists of an input



5.4.3.6 Current Regulator Integral Gain ANN

This network has been designed to select the value of i(ki). It consists of an input




Figure 5.13: Current regulator proportional gain ANN

Figure 5.14: Current regulator integral gain ANN


5.4.3.7 PLL Proportional Gain ANN

This network has been designed to select the value of PLL(kp). It consists of an



Figure 5.15: PLL proportional gain ANN

5.4.3.8 PLL Integral Gain ANN

This network has been designed to select the value of PLL(ki). It consists of an




Figure 5.16: PLL integral gain ANN

5.4.4 Designing and Programming

Generally, programming ANNs follow a number of systemic procedures:

• Collecting of data from the PSO: the values of kp and ki for each case of sag

and swell events using PSO.

• Data reprocessing in this step the data is normalized and the missing data are

replaced by average of neighboring values.

• Building the network: in this step specification of hidden layers, neurons in

each layer, transfer function in each layer, training function, weight/bias learn-

ing and performance function.

• Training the network during the training process, weights are adjusted in order

to make the actual outputs (predicated) close to the measured outputs of the

network.


• Testing the network by used the targets output in simulation model.

Figure5.17 describe the Basic flowchart for designing ANN Model.

Figure 5.17: Basic flowchart for designing ANNs model

5.5 Total Harmonic Distortion

5.5.1 Harmonics Formulation

The harmonic distortion may be present by Foureier series as given in Eq. (5.15) [39]

V (t) = V0 +∞∑n=1

√2Vn cos(n2πf0t+ φn) (5.15)

Where:

V0: is the value of the dc component.


Vn: RMS value of nth harmonic component

f0: fundemetal frequency(50Hz)

φn: the phase angle of the nth harmonics component.

5.5.2 Measurement of THD% In Distribution System

There are many indication for harmonic distortion in the system appear as harmonic

spectrum, RMS voltage and current level. THD giving the power of all harmonics

in relation of fundemental frequency as given in Eq. (5.16) [39]

THD% =

√∑nn=2 V

2n

V1∗ 100% (5.16)

Where:

Vn: the RMS value of the harmonic components of order n of the voltage.

V1: the RMS value of the fundemental component of the voltage.

The THD% is measured for each case of all events (appendix C). MATLAB code

is built for this purpose using the concept of discrete fourier transform(appendix A).

IEEE standard 519-1992 [39] provides the limits of THD% for normal operation.

Table. 5.4 shows the IEEE standard limits for THD%.

Table 5.4: Voltage distortion according to IEEE 519-1992 [39]

Bus Voltage(kV) THD%69 and bellow 5.069 through 161 2.5161 and above 1.5

Chapter 6

Results and Conclusions

6.1 Results and Discussion

6.1.1 Creation of Sag and Swell Events on Distribution Bus

A wide range of extreme sag(0.7pu) and extreme swell(1.3pu) are created on dis-

tributed bus. This may represent the different disturbances occur on distribution

system such as faults or over voltage due switching capacitor bank and other elec-

trical disturbance may occur. Figure 6.1 and Figure 6.2 show the extreme sag and

extreme swell respectively.

55

CHAPTER 6. RESULTS AND CONCLUSIONS 56

Figure 6.1: Extreme sag event 0.7pu

Figure 6.2: Extreme swell event 1.3pu

These events are simulated by network model using MATLAB/Simulink. The

deviation of voltage on distribution bus is considered to be around (0.05)pu. DSTAT-

COM controller gains are obtained for each created event using PSO algorithm as

well as the percentage error is also calculated according to the objective function of

the optimization function. Table. 6.1 and table. 6.2 shows the results of DSTAT-

COM controller gains for the main steps disturbances of swell and sag voltage events

respectively (see all data base for created events and optimal values for controller

gains using PSO in appendix C).

The simulation results of proposed network under extreme sag and extreme swell


Table 6.1: Optimal values of DSTATCOM controller gains during main steps ofswell events using PSO

Event Vac(kp) Vac(ki) Vdc(kp) Vdc(ki) i(kp) i(ki) PLL(kp) PLL(ki) error%

1.30pu 10 10 10 20 2.75 415 1170 3000 4.0009

1.25pu 7.43 3500 1e-4 1.23e1 2.17 1500 1500 10 4.580

1.20pu 3.82 1.98e2 6.22 1.55e1 2.05 4.36e2 1.38e3 2.64e3 5.457

1.15pu 10 10 1e1 1e-2 5 1.28e3 1.50e3 3000 8.866

1.10pu 10 10 10 20 5 1500 1.8e3 3000 5.925

Table 6.2: Optimal values of DSTATCOM controller gains during main steps of sagevents using PSO


0.9pu 1e-2 10 1e-4 1e-2 1 1e2 2.1e2 2.38e1 9.989

0.85pu 1e-2 3500 10 1.67 1 1.17e3 9.88e2 3000 9.552

0.80pu 10 3500 1e-4 20 5 7.01e2 1.13e3 3.24e2 9.141

0.75pu 10 3500 1e-4 20 1 1.37e3 6.59e2 10 9.171

0.70pu 10 10 1e-4 20 1 100 4.55e1 10 0.999

conditions without using DSTATCOM is showed in Figure 6.3.

Figure 6.3: Extreme sag and extreme swell events


6.1.2 Selection of Controller Gains Using PSO

The controllers gains of DSTATCOM are selected and optimized according to the

objective function which is minimization the RMSE between the measured bus volt-

age and the reference value (1pu). The sag events via error is drawn as in Figure6.4

while for swell events via error drawn in Figure 6.5. From the charts of sag and

swell via percentage error, it clear that the worst case of error is below 10% which

acceptable referring to IEEE Standards.

Figure 6.4: Sag events via error%

Figure 6.5: Swell events via error%


6.1.3 Control DSTATCOM using ANNs

PSO is used as a tool to obtain the optimum values of controller gains. Then data

processing is done and nearly 70% of these data is used to train eight networks that

control DSTATCOM as mention in table. 5.3. Values for controller gains is applied

to ANNs in training phase, to obtain a set of weights that will minimize the error

function in competitive time. ANN logarithm is used for practice the controller

and select the gains constants according to activations elements. The controller

constants is used as inputs to train the eight controller constant networks.

Figure 6.6 and Figure 6.7 are show the training, validation, and test errors

used to check the progress of training for eight trained networks. These results

are reasonable since the test set error and the validation set error have similar

characteristics, it does not appear that any significant over fitting has occurred.

The best validation performance for chosen network gave the overall minimum of

system, are reported as given in table. 6.3 for different trained neural networks.


Figure 6.6: Training validation and test curves for neural networks (1-4)


Figure 6.7: Training validation and test curves for neural networks (5-8)

Figure 6.8 and Figure 6.9 describe the estimation function by the ANN. The

network outputs are plotted versus the targets as open circles. The best linear fit is

indicated by dashed line. The perfect fit( the output equal to target) is shown by

solid line. The output seem to track the targets perfectly with regression as reported

in table. 6.3. The R value is an indication of the relationship between outputs and


targets. If R is close to zero, then there is no linear relationship between outputs

and targets, but the research results showed that R value is greater than 0.75 this

indication to the acceptable relationship between outputs and targets for different

neural trained networks.

Figure 6.8: Estimation and fitting for neural networks (1-4)


Figure 6.9: Estimation and fitting for neural networks (5-8)


Table 6.3: Evaluation of MSE and fitting function R for eight neural networks

NeuralNetwork

numberof epoch

MSE R

Network1 7 0.69845 0.81662Network2 5 1.2963 0.85723Network3 12 1.2311 0.75731Network4 2 0.69725 0.85287Network5 2 2.3703 0.81662Network6 1 1.1434 0.85723Network7 14 1.3516 0.75731Network8 4 3.8998 0.85287

6.1.4 Effect of ANNs Controller on Performance of DSTAT-

COM for Improve the Voltage Profile

ANN logarithm is used for practice the controller and select the its gains constants

according to activations elements. ANNs is quite easy but requires a large a mount

of data related to the input parameters to cover a wide range of possibilities to

predict the output. The algorithms is employed using nearly 70% of data obtained

from PSO. Table. 6.4 shows the results and controller gains for extreme swell and

extreme sag by using developed controller of DSTATCOM.

The DSTATCOM is improve the voltage profile in both cases (swell and sag) to

nearly 1.05pu which acceptable according to IEEE Std.1159 as shown in Figure 6.10

and Figure 6.11 respectively. After the ANNs is applied on cases between sag and

swell for controllers gains obtained by PSO The DSTATCOM improve the voltage

profile around (1.02pu) and the THD is decrease more than using only PSO. This

lead to use ANN as an AI technique to control DSTATCOM for voltage profile

improvement on distribution system for the goal of real time control rather than


Table 6.4: Voltage profile improvement using developed DSTATCOM controller

Event Controllers kp ki THD% Voltageprofile im-provement

Swell(1.3)pu

Vac Regulator Gains 7.429 3500

3.26% 1.02puVdc Regulator Gains 1e-4 1.225e+1Current Regulator Gains 2.176e+4 1500PLL Regulator Gains 1500 10

Sag(0.7)pu

Vac Gegulator Gains 10 3500

9.23% 1.07puVdc Regulator Gains 1e-1 1e-2Current Regulator Gains 5 1500PLL Regulator Gains 200 500

traditional off line control.

From the above results using combined AI techniques with optimization topology

give more accuracy for power system this lead to increase the reliability of power

system for solve PQ problems in distributions system especially when a nonlinear

and sensitive loads connect on the end user side. Also the controller is examined

for improving the voltage profile for continuous extreme sag and extreme swell, the

results proved the robust and efficient of developed controller as shown in Figure

6.12 and Figure 6.13 for rms voltage and phase voltage respectively.

6.1.5 THD Measurement

MATLAB code is written for purpose of evaluation THD based on FFT, which is

calculated during all created events. It’s observed that the THD in sag case is greater

than in swell case this due in this mode DSTATCOM supplied reactive power to

the grid. The THD via voltage events are plotted, the results shown in figure 6.14


Figure 6.10: Voltage profile improvement for extreme swell event

Figure 6.11: Voltage profile improvement for extreme sag event

and figure 6.15 for swell and sag respectively. Frequency spectrum is plotted using

MATLAB code(appendix A) as shown in figure 6.16 and figure 6.17 for extreme sag

and extreme swell respectively.


Figure 6.12: rms voltage profile improvement using DSTATCOM tuned by devel-oped DSTATCOM controller using AI

Figure 6.13: Phase voltage profile improvement using DSTATCOM tuned by de-veloped DSTATCOM controller using AI


Figure 6.14: THD for swell events

Figure 6.15: THD for sag events

The results proved the importance of using AI in controller design of DSTAT-

COM. In This work the voltage profile is improved for extreme sag and extreme

swell events real time. On the other hand, THD decrease more that other methods

as mentioned in literature [43]- [45], but still needed more improvement may be

done by modifying the filter design of DSTATCOM.


Figure 6.16: Frequency spectrum for extreme sag event

Figure 6.17: Frequency spectrum for extreme swell event


6.2 Conclusions

The main goal of this thesis is to improve the voltage profile of distribution system

using DSTATCOM, which controlled by AI Technique. The PSO is used as a tool

for obtained the optimum values of DSTATCOM controller gains and then ANN is

used as AI technique for real time control of DSTATCOM. The work proved the

efficient of using ANN as an AI technique for tuning PI controller of DSTATCOM

in order to improve the voltage profile on distribution system. The DSTATCOM

with developed controller validated the importance of AI methods such as ANN.

Using such this developed controller may increase stability and reliability of power

system and hence the network convert to smart network.

Developed controller using ANN supports DSTATCOM with robust tuning sys-

tem, since its responds to different real time voltage disturbances that may occur

in distribution system with efficient enhancement for voltage profile compared with

conventional control strategies.

The simulation results proved the efficient of suggested method for controlling

DSTATCOM in order to improve voltage profile on distribution system.

6.3 Future Works

Although this work covered using ANN as an AI technique tuning PI controller of

DSTATCOM for purpose of improve the voltage profile on distribution system ad-

ditional work is vital for future research.

It is interesting to extend this work by employ DSTATCOM with improved AI

control technique on the real distribution system with various load conditions and


evaluate the efficient of using developed controller of DSTATCOM to improve dif-

ferent PQ issues.

The operation of DSTATCOM is mainly dependent on an important factor which

is dc storage element, this factor is affected on the voltage regulation on distribution

system. So it may be one of considerable topics for researchers to study the facilities

to combine the RESS as the main supply of energy for DSTATCOM as well as

determine the optimum and most efficient dc storage device for DSTATCOM design.

The filter design of DSTATCOM have a great challenges to decrease the THD%.

Bibliography

[1] Gandoman, F.H., Ahmadi, A., Sharaf, A.M., Siano, P., Pou, J., Hredzak, B. and

Agelidis, V.G., 2018. Review of FACTS technologies and applications for power

quality in smart grids with renewable energy systems. Renewable and sustainable

energy reviews, 82, pp.502-514.

[2] Sharaf, A.M. and Gandoman, F.H., 2015, August. A switched hybrid filter-

DVS/green plug for smart grid nonlinear loads. In Smart Energy Grid Engineer-

ing (SEGE), 2015 IEEE International Conference on (pp. 1-6). IEEE.

[3] Farhoodnea, M., Mohamed, A. and Shareef, H., 2014, May. A comparative study

on the performance of custom power devices for power quality improvement. In

Innovative Smart Grid Technologies-Asia (ISGT Asia), 2014 IEEE (pp. 153-157).

IEEE.

[4] Ghosh, A. and Ledwich, G., 2012. Power quality enhancement using custom

power devices. Springer Science & Business Media.

[5] Short, T.A., 2005. Distribution reliability and power quality. Crc Press.

72

BIBLIOGRAPHY 73

[6] Kalair, A., Abas, N., Kalair, A.R., Saleem, Z. and Khan, N., 2017. Review of har-

monic analysis, modeling and mitigation techniques. Renewable and Sustainable

Energy Reviews, 78, pp.1152-1187.

[7] Kalair, A., Abas, N., Kalair, A.R., Saleem, Z. and Khan, N., 2017. Review of har-

monic analysis, modeling and mitigation techniques. Renewable and Sustainable

Energy Reviews, 78, pp.1152-1187.

[8] Singh, B. and Solanki, J., 2009. A comparison of control algorithms for DSTAT-

COM. IEEE transactions on Industrial Electronics, 56(7), pp.2738-2745.

[9] Hosseini, S.H., Nazarloo, A. and Babaei, E., 2010, August. Application of D-

STATCOM to improve distribution system performance with balanced and un-

balanced fault conditions. In Electric Power and Energy Conference (EPEC),

2010 IEEE (pp. 1-6). IEEE.

[10] Roy, N.K., Hossain, M.J. and Pota, H.R., 2011, July. Voltage profile improve-

ment for distributed wind generation using D-STATCOM. In Power and Energy

Society General Meeting, 2011 IEEE (pp. 1-6). IEEE.

[11] Singh, B., Adya, A., Mittal, A.P. and Gupta, J.R.P., 2005. Power quality en-

hancement with DSTATCOM for small isolated alternator feeding distribution

system. In Power Electronics and Drives Systems, 2005. PEDS 2005. Interna-

tional Conference on (Vol. 1, pp. 274-279). IEEE.

[12] Ipinnimo, O., Chowdhury, S., Chowdhury, S.P. and Mitra, J., 2013. A review

of voltage dip mitigation techniques with distributed generation in electricity

networks. Electric Power Systems Research, 103, pp.28-36.

BIBLIOGRAPHY 74

[13] Mahyavanshi, A., Mulla, M.A. and Chudamani, R., 2012. Reactive power com-

pensation by controlling the DSTATCOM. International Journal of Emerging

Technology and Advanced Engineering, 2(11), pp.212-218.

[14] Singare, A.R., Bagekar, S.D., Raut, D.B. and Naranje, S.M., 2018. Compensa-

tion of Voltage Swell by Using STATCOM.

[15] Kumar, M., Swarnkar, A., Gupta, N. and Niazi, K.R., 2017. Design and op-

eration of DSTATCOM for power quality improvement in distribution systems.

The Journal of Engineering, 2017(13), pp.2328-2333.

[16] Kullan, M., Muthu, R., Mervin, J.B. and Subramanian, V., 2016. Design of

DSTATCOM Controller for Compensating Unbalances. Circuits and Systems,

7(09), p.2362.

[17] Afonso, J.L., Freitas, M.S. and Martins, J.S., 2003, June. pq Theory power

components calculations. In Industrial Electronics, 2003. ISIE’03. 2003 IEEE

International Symposium on (Vol. 1, pp. 385-390). IEEE.

[18] Etxeberria, A., 2012. Microgrid hybrid energy storage and control using a three-

level NPC converter (Doctoral dissertation, Bordeaux 1).

[19] CChen, C.L., Wang, Y., Lai, J.S., Lee, Y.S. and Martin, D., 2010. Design of

parallel inverters for smooth mode transfer microgrid applications. IEEE Trans-

actions on Power Electronics, 25(1), pp.6-15.

[20] Ramesh, G.B. and Pruthviraja, L., 2016. Applications of D-STATCOM for

Power Quality Improvement in Distribution System.

BIBLIOGRAPHY 75

[21] Molina, M.G. and Mercado, P.E., 2006. Control design and simulation

of DSTATCOM with energy storage for power quality improvements. Proc.

Ieee/Pes Trans. Distrib. La, pp.01-07.

[22] Liu, C.H. and Hsu, Y.Y., 2010. Design of a self-tuning PI controller for a

STATCOM using particle swarm optimization. IEEE Transactions on industrial

Electronics, 57(2), pp.702-715..

[23] Ramya, G., Ganapathy, V. and Suresh, P., 2017. Power quality improvement

using multi-level inverter based DVR and DSTATCOM using neuro-fuzzy con-

troller. International journal of power electronics and drive systems (IJPEDS),

8(1), pp.316-324.

[24] Abdelsalam, A.A., Desouki, M.E. and Sharaf, A.M., 2013. Power quality im-

provement using FACTS power filter compensation scheme. Journal of Electrical

Systems, 9(1), pp.73-83.

[25] Kow, K.W., Wong, Y.W., Rajkumar, R.K. and Rajkumar, R.K., 2016. A review

on performance of artificial intelligence and conventional method in mitigating

PV grid-tied related power quality events. Renewable and Sustainable Energy

Reviews, 56, pp.334-346.

[26] Firouzjaee, H.A., Kordestani, J.K. and Meybodi, M.R., 2017. Cuckoo search

with composite flight operator for numerical optimization problems and its ap-

plication in tunneling. Engineering Optimization, 49(4), pp.597-616.

[27] Metropolis, N., Rosenbluth, A.W., Rosenbluth, M.N., Teller, A.H. and Teller,

E., 1953. Equation of state calculations by fast computing machines. The journal

of chemical physics, 21(6), pp.1087-1092.

BIBLIOGRAPHY 76

[28] Ghosh, S. and Sherpa, K.S., 2008. An efficient method for load-flow solution

of radial distribution networks. Int. J. Elect. Power Energy Syst. Eng, 1(2),

pp.108-115.

[29] Eswaran, T. and Kumar, V.S., 2017. Particle swarm optimization (PSO)-based

tuning technique for PI controller for management of a distributed static syn-

chronous compensator (DSTATCOM) for improved dynamic response and power

quality. Journal of Applied Research and Technology, 15(2), pp.173-189.

[30] Chattopadhyay, S., Mitra, M. and Sengupta, S., 2011. Electric power quality.

In Electric Power Quality (pp. 5-12). Springer, Dordrecht.

[31] Hussain, S.A., Zaro, F.R. and Abido, M.A., 2013, February. Implementation

of quadrature based RMS calculation on real-time power monitoring systems.

In Power and Energy Conference at Illinois (PECI), 2013 IEEE (pp. 213-217).

IEEE.

[32] Kumar, S.R. and Nagaraju, S.S., 2007. Simulation of D-STATCOM and DVR

in Power Systems. ARPN Journal of Engineering and Applied Sciences, 2(3),

pp.7-13.

[33] Ahmadi, H., Mart, J.R. and Dommel, H.W., 2015. A framework for volt-VAR

optimization in distribution systems. IEEE Transactions on Smart Grid, 6(3),

pp.1473-1483.

[34] Labeeb, M. and Lathika, B.S., 2011, September. Design and analysis of DSTAT-

COM using SRFT and ANN-fuzzy based control for power quality improvement.

In 2011 IEEE Recent Advances in Intelligent Computational Systems (pp. 274-

279). IEEE.

BIBLIOGRAPHY 77

[35] Bashir, Z.A. and El-Hawary, M.E., 2009. Applying wavelets to short-term load

forecasting using PSO-based neural networks. IEEE transactions on power sys-

tems, 24(1), pp.20-27.

[36] Mohaghegi, S., del Valle, Y., Venayagamoorthy, G.K. and Harley, R.G., 2005,

June. A comparison of PSO and backpropagation for training RBF neural net-

works for identification of a power system with STATCOM. In Swarm Intelligence

Symposium, 2005. SIS 2005. Proceedings 2005 IEEE (pp. 381-384). IEEE.

[37] Mallick, R.K., Mishra, A. and Sinha, S., 2015. Improvement of power qual-

ity in DFIG based wind farm using optimized DSTATCOM. In International

Conference (pp. 1-7).

[38] Nabisha, A. and Joseph, X.F., 2017. Power Quality Enriched Wind Energy

System using DSTATCOM based on PID-ANN Controller. American-Eurasian

Journal of Scientific Research, 12(5), pp.271-284.

[39] Masdi, H., Mariun, N., Bashi, S.M., Mohamed, A. and Yusuf, S., 2009. Con-

struction of a prototype D-Statcom for voltage sag mitigation. European Journal

of Scientific Research, 30(1), pp.112-127.

[40] Panda, S. and Padhy, N.P., 2008. Comparison of particle swarm optimization

and genetic algorithm for FACTS-based controller design. Applied soft comput-

ing, 8(4), pp.1418-1427.

[41] Babaei, M., Jafari-Marandi, R., Abdelwahed, S. and Smith, B., 2017, February.

A simulated Annealing-based optimal design of STATCOM under unbalanced

conditions and faults. In Power and Energy Conference at Illinois (PECI), 2017

IEEE (pp. 1-5). IEEE.

BIBLIOGRAPHY 78

[42] Niknam, T., Firouzi, B.B. and Ostadi, A., 2010. A new fuzzy adaptive particle

swarm optimization for daily Volt/Var control in distribution networks consid-

ering distributed generators. Applied Energy, 87(6), pp.1919-1928.

[43] Alam, M.N., Das, B. and Pant, V., 2015. A comparative study of metaheuristic

optimization approaches for directional overcurrent relays coordination. Electric

Power Systems Research, 128, pp.39-52.

[44] Alam, M.N., 2016. Particle swarm optimization: Algorithm and its codes in

matlab. ResearchGate, pp.1-10.

[45] Wang, L., Lam, C.S. and Wong, M.C., 2016. A hybrid-STATCOM with wide

compensation range and low DC-link voltage. IEEE Transactions on Industrial

Electronics, 63(6), pp.3333-3343

[46] Ahmadi, H., Mart, J.R. and Dommel, H.W., 2015. A framework for volt-VAR

optimization in distribution systems. IEEE Transactions on Smart Grid, 6(3),

pp.1473-1483.

[47] Labeeb, M. and Lathika, B.S., 2011, September. Design and analysis of DSTAT-

COM using SRFT and ANN-fuzzy based control for power quality improvement.

In Recent Advances in Intelligent Computational Systems (RAICS), 2011 IEEE

(pp. 274-279). IEEE.

[48] Li, Y., 2017. Ensemble bias-correction based state of charge estimation of

lithium-Ion batteries.

BIBLIOGRAPHY 79

[49] Panda, S. and Padhy, N.P., 2008. Comparison of particle swarm optimization

and genetic algorithm for FACTS-based controller design. Applied soft comput-

ing, 8(4), pp.1418-1427.

[50] Beiki, E.B., Arslan, O., Turan, O. and ler, A.I., 2016. An artificial neural net-

work based decision support system for energy efficient ship operations. Com-

puters Operations Research, 66, pp.393-401.

[51] Agatonovic-Kustrin, S. and Beresford, R., 2000. Basic concepts of artificial

neural network (ANN) modeling and its application in pharmaceutical research.

Journal of pharmaceutical and biomedical analysis, 22(5), pp.717-727.

Appendix A

MATLAB Codes

\\sectionPSO Alogorithm

%% -Help psoToolbox V1.0

% psoToolbox provides an interective GUI based Toolbox to solve

% optimization problems using particle swarm optimization.

% Creat a fitness function in M-file.

% Inputs:

% Function : Function handle of fitness function.

% Nvars : Number of variable to be optimized.

% LB : Lower Bound of Nvars (1 X Nvars)

% UB : Upper Bound of Nvars (1 X Nvars)

% Parameters:

% C1 : Cognative Attraction

% C2 : Social Attraction

% W : Inertial

% Population Size : Number of Swarms

% Max Iterations : Maximum number of epochs.

% Click on " RUN PSO " button to start PSO search. You will get the out put

% at Edit box below the axes.

%

% To solve problem on command prompt use "pso.m".

%% Figure and Back ground

set(0,’units’,’pixel’);

screensize = get(0,’screensize’);

height = 600;

width = 800;

80

APPENDIX A. MATLAB CODES 81

basecolor = [rand 1 rand]./2;

imdata = creatback(height,width,basecolor);

fig_position = [(screensize(3)/2)-(width/2) (screensize(4)/2)-(height/2) width height];

handle.fig = figure(’Name’,’PSOToolbox’,...

’units’,’pixel’,...

’position’,fig_position,...

’numbertitle’,’off’,...

’menubar’,’none’,...

’resize’,’off’);

’foregroundcolor’,’w’,...

’horizontalalignment’,’left’);

handle.text12 = text(35,160,’Nvars : ’,...

’fontsize’,14,...

’fontname’,’cambria’,...

’color’,’w’);

handle.edit12 = uicontrol(’style’,’edit’,...

’parent’,handle.fig,...

’position’,[132 428 150 25],...

’backgroundcolor’,basecolor*(height-428)/height,...





handle.text13 = text(35,210,’Lower Bound : ’,...






’position’,[162 378 120 25],...






handle.text14 = text(35,260,’Upper Bound : ’,...






’position’,[162 328 120 25],...









’position’,[186 223 90 25],...






handle.text23 = text(35,415,’W : ’,...






’position’,[72 173 90 25],...






handle.text24 = text(35,465,’Population Size : ’,...

’color’,basecolor*(height-380)/height,...

’xcolor’,’w’,...

’ycolor’,’w’);

xlabel(’Iteration’);

ylabel(’F(x)’);

h=title(’Convergence of F(x)’);

set(h,’color’,’w’);

grid on

’callback’,@push1_callback);

%% Background Image

function cdata = creatback(height,width,basecolor)

cdata = zeros(height,width,3);

[height , ~, page] = size(cdata);

for i = 1:height

color = basecolor.*(i/height);

for j = 1:page

cdata(i,:,j) = color(j);

end

end


end

%% Default

set(handle.edit21,’string’,’1.2’);%C1

dit12,’string’));% nvars

LB = str2num(get(handle.edit13,’string’));% LB

UB = str2num(get(handle.edit14,’string’));%UB

if isempty(ObjectiveFunction)

errordlg(’PS0 : Define Objective Function’);

else

ObjectiveFunction=str2func(ObjectiveFunction);

end

if ~isequal(Nvars,size(LB,2),size(UB,2))

errordlg(’PSO : Invalid Arguments : isequal(Nvars,size(LB,2),size(UB,2)) should be true ’);

end

str1 = get(handle.edit3,’string’);

str2 = sprintf(’\nRunning Optimization....’);

set(handle.edit3,’string’,str);

% PSO

CurrentPosition = zeros(N,Nvars); % Initial Position

for i = 1:Nvars

CurrentPosition(:,i) = random(’unif’,LB(i),UB(i),N,1);

end

Velocity = W.*rand(N,Nvars) ; % Initial Velocity

% Evalute Initial Position

CurrentFitness = zeros(N,1); % Fitness Value

for i = 1:N

assignin(’base’,’x’,CurrentPosition(i,:));

[CurrentFitness(i) erorr1] = ObjectiveFunction(CurrentPosition(i,:));

end

for i = 1:Nvars

indexes = CurrentPosition(:,i) < LB(i).*ones(N,1);

CurrentPosition(indexes,i) = LB(i);

indexes = CurrentPosition(:,i) > UB(i).*ones(N,1);

CurrentPosition(indexes,i) = UB(i);

end

if (Iter==1)||(rem(Iter,1)==0)

p;

xlabel(’Iteration’);

ylabel(’F(x)’);

h = title([’Best Value : ’ num2str(GlobalBestFitness)]);

set(h,’color’,’w’);

grid on


end

end

str3= sprintf(’\nOptimization Complete.’);


str4 = sprintf(’\nF(x) = %d’,GlobalBestFitness);

str5 = sprintf(’\nx = %d’,GlobalBestPosition(1,:));


clear(’str’);

end

end

-------------------------------------------------------------

\\ Objective function code

function [F,erorr1]= tracklnoise(z)

% Track the output of optsim to a signal of 1

% Variables a1 and a2 are shared with RUNTRACKLSQ

Kd =0 ;

% Compute function valuew

simout=sim(’power_dstatcom_pwm2swell844’);

m=length(voutrms(20000:end,2));

t=0:5e-6:5e-6*(m-1);

signal=sin(2*pi*50*t+asin(three_pahe(20000,2)));

e=sqrt(sum((signal-three_pahe(20000,2)).^2)/length(voutrms(20000:end,2))) ; % compute the error

erorr1=100*sum((1-voutrms(20000:end,2)).^2)/length(voutrms(20000:end,2)) ; % compute the error

assignin(’base’,’erorr1’,erorr1);

F= erorr1;

end

------------------------------------------------

\\ NNs Code


Target=table(:,8)’;

Input=x;

net8= newff(Input,Target, [3], ’tansig’,’tansig’,’trainlm’, ’learngdm’,’mse’);

% Define learning parameters

net.trainParam.epochs=1000000; % Maximum number of epochs to train

net.trainParam.goal = 1e-12; %Performance goal

net.trainParam.lr=0.15;%Learning rate

net.trainParam.mc=0.6;%Momentum constant

net.trainParam.lr_inc=1.05; %Ratio to increase learning rate

net.trainParam.lr_dec=0.7; % Ratio to decrease learning rate%

net.trainParam.max_fail=1000; %Maximum validation failures

net.trainParam.max_perf_inc=1.04; % Maximum performance increase%

net.trainParam.min_grad=1e-10; %Minimum performance gradient

net.trainParam.show=300; %Epochs between displays (NaN for no displays)

net.trainParam.time=inf;%Maximum time to train in seconds

find_gain_constants = train(net8,Input,Target);

--------------------------------------------------

\\ THD Code

clear all;close all; clc;

rms_value(k,:)=voutrms(:,2)’;

erorr(k)=100*sum((1-voutrms(20000:end,2)).^2)/length(voutrms(20000:end,2)) ; % compute the error

signal=three_pahe(1900001:2000001,2);

Y=fft(signal);

L=length(signal);

P1=abs(Y/L);

P2=P1(1:L/2+1);

P2(2:end-1)= 2*P2(2:end-1);

f=(0:L/2)/(L*Ts);

P3=sqrt(sum((P2.^2))-max(P2)^2)/max(P2);

stem(f,P2)

THD(k)=P3;

k=k+1;

end

Appendix B

Real feeder given by JDECO

1. Abu Mashaal given by JDECO as a case study

Figure B.1: Abu Mashaal case study given by JDECO

86

APPENDIX B. REAL FEEDER GIVEN BY JDECO 87

2. Load and voltage profile

Figure B.2: Load and Voltage Profile

APPENDIX B. REAL FEEDER GIVEN BY JDECO 88

3. Data for line feeder (Der Abu Mashaal)

Figure B.3: Der Abu Mshaal Data Feeder

Appendix C

Optimum values of controllergains for swell and sag voltageevents

1. DSTATCOM Controller Gains For Swell Events Using PSO

2. DSTATCOM Controller Gains For Sag Events Using PSO

89

APPENDIX C. OPTIMUMVALUES OF CONTROLLERGAINS FOR SWELL AND SAGVOLTAGE EVENTS 90

Table C.1: DSTATCOM Controller Gains For Swell Events Using PSO


1.30pu 10 10 10 20 2.75 415 1170 3000 4.0009

1.295pu 0.834 3100 5.73 18.8 5 1050 1290 2310 2.960

1.29pu 10 3500 1e-4 20 5 4.08e2 1.42e3 1.11e2 3.599

1.285pu 10 10 10 20 4.97 1.06e3 1.42e3 3000 4.440

1.28pu 2.48 10 1e-4 1.72e1 4.71 1e2 1500 10 3.517

1.275pu 7.57 3500 0.46 0.92e-1 5 1500 1.25e3 1.91e3 4.880

1.27pu 3.19 3270 10 20 5 100 1.23e3 5.77e2 3.440

1.265pu 1e-2 2.55e3 10 20 5 1.17e2 1.26e3 10 4.997

1.26pu 10 3500 10 1e-2 5 1.10e2 1500 10 4.339

1.255pu 10 3500 1e-4 1e-2 5 1500 1.25e3 10 3.950

1.25pu 7.43 3500 1e-4 1.23e1 2.17 1500 1500 10 4.580

1.245pu 4.35 4.71e2 3.97 20 2.50 0.44e2 1.33e3 1.57e3 7.553

1.24pu 4.15 4.31e2 3.57 20 3.57 4.39e2 1e3 1.27e3 3.353

1.235pu 2.26 3.31e2 2.97 20 2.40 3.36e2 1.02e3 1.57e3 5.335

1.23pu 6.79 4.61e2 10 3.44 5 2.51e2 1.27e3 2.84e3 4.813

1.225pu 4.14 3.50e3 5.02 3.53 5 1.78e2 1.42e3 8.26e2 6.487

1.22pu 4.13 3500 5.02 3.53 5 1.76e2 1.22e3 8.48e2 3.990

1.215pu 1e-2 3500 10 20 5 100 1.17e3 10 3.915

1.21pu 1e-2 3500 10 20 5 100 1.27e3 10 4.451

1.205pu 10 10 3.61 1e-2 5 1500 1.31e3 2.70e2 4.640

1.20pu 3.82 1.98e2 6.22 1.55e1 2.05 4.36e2 1.38e3 2.64e3 5.457

1.195pu 1e1 3500 1e-4 1e-2 5 1500 1.37e4 10 3.852

1.19pu 10 3500 1e-4 1e-2 5 1500 1.36e3 10 7.709

1.185pu 4.33 1.31e2 589 1.2 4.59 6.24e2 1.32e3 2.52e3 7.604

1.18pu 8.54 2.19e3 2.25 1.71e1 3.38 6.93e2 1.27e3 1.27e3 7.227

1.175pu 8.54 2.19e3 2.25 1.71e1 3.38 6.93e2 1.27e3 1.27e3 7.227

1.17pu 2.41 2.83e3 1.89 4.85e-1 3.71 3.09e2 1.43e3 2.96e3 9.651

1.165pu 10 10 5.49 1.06e1 5 8.68e2 1.50e3 1.19e3 4.543

1.16pu 10 10 5.49 1.06e1 5 8.68e2 1.50e3 1.19e3 4.543

1.155pu 10 3500 1e-4 2e1 5 1500 1.50e3 10 7.137

1.15pu 10 10 1e1 1e-2 5 1.28e3 1.50e3 3000 8.866

1.145pu 10 10 1e-4 1e-2 1 1500 1.5e3 10 6.974

1.14pu 1e-2 3500 1e-4 1e-2 5 100 1.26e3 3000 5.774

1.135pu 10 3500 1e-4 2e1 5 1500 1.50e3 10 4.436

1.13pu 1.28 2.3e3 3.11 1.81e1 4.25 8.3e2 7.32e1 2e4 9.989

1.125pu 4.13 1.06e3 7.19 1.96 4 1.07e3 1.22e3 1.5e3 5.646

1.12pu 10 3500 1e-4 1e-2 5 9.74e2 1.50e3 5.58e2 7.713

1.115pu 10 2.77e2 1e1 1.01e1 3.49 100 1.50e3 3000 7.620

1.11pu 5.02 7.37e2 5.18 1.02e1 2.83 100 1.24e3 2.479e3 6.403

1.105pu 4.33 2.05e3 4 1.25e1 3.91 8.4e2 1.81e3 1.18e3 7.569

1.10pu 10 10 10 20 5 1500 1.8e3 3000 5.925

APPENDIX C. OPTIMUMVALUES OF CONTROLLERGAINS FOR SWELL AND SAGVOLTAGE EVENTS 91

Table C.2: DSTATCOM Controller Gains For Sag Events Using PSO


0.9pu 1e-2 10 1e-4 1e-2 1 1e2 2.1e2 2.38e1 9.989

0.895pu 1e1 1.22e3 1e1 1e-2 1 1.5e3 9.2e2 2.77e3 8.144

0.89pu 4.93 3.47e3 2.65 1.87e1 9.9e-1 1.86e2 2.21e2 1.06e3 9.760

0.885pu 4.86 2.45e3 9.13 8.08 1.92 1.21e3 9.77e2 2.01e3 8.311

0.88pu 4.86 2.45e3 9.13 8.08 1.92 1.21e3 9.77e2 2.01e3 8.311

0.875pu 10 3500 2.75 1.39e1 1 1500 9.95e2 10 9.856

0.87pu 1e-2 3500 10 20 1 100 9.45e2 10 9.656

0.865pu 10 1.18e3 10 1e-2 1 1e2 9.35e2 3000 6.93

0.86pu 1e-2 3500 10 20 1 1500 9.75e2 10 7.004

0.855pu 9.9 2.05e3 2.87 9.65 4.14 4.01e2 9.72e2 2.18e3 9.784

0.85pu 1e-2 3500 10 1.67 1 1.17e3 9.88e2 3000 9.552

0.8455pu 1e-2 10 1e-4 1e-2 1 100 1.03e3 10 6.781

0.84pu 10 2.02e3 10 20 1 1500 1.08e3 10 7.775

0.835pu 10 2.02e3 10 20 1 1500 1.08e3 10 7.775

0.83pu 8.97 3500 10 20 1 1500 9.36e2 9.36e2 8.5223

0.825pu 8.97 3500 10 20 1 1500 9.36e2 9.36e2 8.5223

0.82pu 1e-2 10 10 1e-2 1 100 9.72e2 3000 6.480

0.815pu 1e-2 10 10 1e-2 1 100 9.72e2 3000 6.4808

0.81pu 7.16 2670 10 1.68e1 2.61 1.28e3 1000 1.1e3 8.310

0.805pu 7.16 2670 10 1.68e1 2.61 1.28e3 1000 1.1e3 8.310

0.80pu 10 3500 1e-4 20 5 7.01e2 1.13e3 3.24e2 9.141

0.795pu 10 2900 1e-4 20 1 100 4.46e1 3000 7.376

0.79pu 10 2900 1e-4 20 1 100 4.46e1 3000 7.376

0.785pu 10 2900 1e-4 20 1 100 4.46e1 3000 7.376

0.78pu 10 2900 1e-4 20 1 100 4.46e1 3000 7.376

0.775pu 10 3500 10 20 1 1500 4.29e1 3000 7.059

0.77pu 10 2630 1e-4 1e-2 1 1500 10 1.37e2 0.641

0.765pu 10 2630 1e-4 1e-2 1 1500 10 1.37e2 0.641

0.76pu 10 3500 1e-4 1e-2 1 100 10 3.26e2 0.555

0.755pu 10 10 10 6.28 1 100 4.4e1 3000 4.833

0.75pu 10 3500 1e-4 20 1 1.37e3 6.59e2 10 9.171

0.745pu 10 10 10 1e-2 1 100 6.71e1 3000 6.041

0.74pu 10 1930 7.14 2.66 1 100 7.47e2 10 2.146

0.735pu 10 1.06e3 10 1e-2 1 1e2 7.30e2 3000 0.251

0.73pu 10 1.06e3 10 1e-2 1 1e2 7.30e2 3000 0.251

0.725pu 10 3500 10 20 1 100 7.30e1 3000 0.184

0.72pu 9.02 6.99e2 1e-4 20 1 8.82e2 10 1.23e2 7.697

0.715pu 1e-2 3500 10 20 1 8.01e2 10 10 0.812

0.71pu 10 10 1e-4 20 1 100 7.61e2 10 9.209

0.705pu 10 3500 10 3.99 1 100 6.71e2 10 8.134

0.70pu 10 10 1e-4 20 1 100 4.55e1 10 0.999

palestine polytechnic university deanship of graduate

Documents